What will my GPA be?
Big data with the help of machine learning and deep learning algorithms helps us to find the underlying patterns in data and discover hidden knowledge, and eventually make better decisions. One of the application areas of big data recently is the higher education, where student data (demographic and grades) is used to understand the problems in students’ difficulties in learning and achievements, to predict student success in courses or programs, and also deal with many more issues in higher education programs. Colleges around the world are using big data to increase retention rates, improve the learning of students, shorten study periods and more. In this project we used student and course grade data, and applied linear and logistic regression, MLP, and Support Vector Machine algorithms, (1) to predict the study length of students, (2) to find the likelihood of student dropout, (3) to predict student and course grades. Our results show that we can predict student dropouts by 89%, student course grades by 84% accuracy.
GitHub
Project Web Page
Video
Project Poster
Project Presentation
Meltem DAŞDEMİR
Efnan GÜLKANAT
Sergen İSPİR
Advisor: Prof. Dr. Erdoğan DOĞDU
Augmented Reality Based Continuous Onboarding Framework
(received TÜBİTAK senior project support)
Augmented reality (AR) is an enhancing area that is getting more and more used in different areas. It becomes a promising technology used in various fields such as education, industry, entertainment, and military. Yet, its usage in software engineering is very limited. The goal of our project is combining augmented reality and software development workflow process and providing continuous onboarding to software practitioners while performing their daily tasks. Software developers should be interactively guided using AR technology on the onboarding process continuously. Software teams may benefit from digitally enhanced working conditions provided using AR. “Augmented Reality Based Continuous Onboarding Framework” is an onboarding system for new-coming software practitioners for getting familiar with the colleagues, company culture and project meetings and information. After researching many different articles and sources, we deduced that there is lack of research about the combination of augmented reality, software engineering disciplines and onboarding processes. In this project conducted by Havelsan Inc, new comers to the company can continuously be onboarded with this application that is run on smart mobile phones and augmented reality based devices.
GitHub
Project Web Page
Video
Project Poster
Project Presentation
İpek OHRİ
Bora ORKUN
H. İrem ÖGE
Advisor: Dr. Murat YILMAZ
VRelax: Investigation of Relaxing
Anxiety is a crucial disease for the people which is not only a problem that decreasing patients’ life quality, but also challenging situation for psychologists who are trying to examine this kind of patients. This project creates an opportunity for psychologists to understand these patients’ stressful moments, at the same time, allows patients to relax themselves on their own. Main purpose of this project is to detect participant’s stress level, and try to reduce stress with relaxing scenarios. Because of the advanced sense of reality that VR headsets provide, VRelax can comfort participants more effectively than trying to comfort the participant via talking with him/her. Project’s main working principle is based on voice analysis process that a virtualized system created in a game engine will use this process to maximize utility, so participant shall be able to interact with the environment with different functional capabilities while analyzing stress. These interactions of participant and scenarios make the created world is more realistic to the participant which with this sense of reality makes reducing of stress relatively easy.
GitHub
Project Web Page
Video
Project Poster
Project Presentation
Görkem NARİNOĞLU
Umut EFİLOĞLU
Advisor: Dr. Faris Serdar TAŞEL
A Serious Game for COE-DAT as a Virtual Policymaking Laboratory
Traditionally, games are considered as a source of entertainment . However, they can be equipped with educational content ; which is known as “serious games” . In this project, serious game theory has been used . Recently, there is a growing interest towards serious games in various fields . Since serious games can be utilized for wide range of opportunities . In this project, we designed a military serious game to provide a virtual policymaking laboratory, which can be used by NATO Centre of Excellence Defence Against Terrorism (COE – DAT) trainees on counter terrorism . The output of this project can be used as a complementary tool to support theoretical counter – terrorism courses.
GitHub
Project Web Page
Video
Project Poster
Project Presentation
Sümeyye TEZALAN
Büşra YILMAZ
Merve YORULMAZ
Advisor: Dr. Murat YILMAZ, Ulaş GÜLEÇ
Gesture Recognizer Strap using Motion Sensor
Today a lot of systems are operated by using the gesture recognition technology in different areas. The improvement in the technology make this topic more popular. The purpose of the project ‘Gesture Recognizer Strap using Motion Sensor’ is to develop an gesture recognizer software, which includes an API to drive the wearable technology, and to develop affordable multipurpose device. In order to reach the aim of the project, an accelerometer, gyroscope, compass sensors will be utilized as these sensors are the most affordable solution for gesture recognition problem. The system will be mounted on a wristband to form a wearable device which tracks the basic movements of arm of the user like linear movements of arm (such as up, down, left, right) or angular movements of the arm.
GitHub
Project Web Page
Video
Project Poster
Project Presentation
Bengi GÜNAY
Advisor: Dr. Faris Serdar TAŞEL
Search by Question
Finding correct answers to queries in search engines is still a problem as most of the search engines do search by keyword based interfaces. In this project, a question and answer type of intelligent search engine prototype is developed by exploring the data collection using interactive filtering to learn the actual intent of the user. Then, search by question finds results of asked 5w1h question in the document collection and list them. Here, IBM Watson tool is used to process the data collection. Search by Question is basis of semantic analysis that try to answer 5w1h questions. The main point here is natural language processing, NLP for short, is a method for to analyze, understand, and make meaning from human language in a smart way. To produce a solution, NLP should be used efficiency and develop an useful algorithm to extract the possible answer(s) in returned data from IBM Watson.
GitHub
Project Web Page
Video
Project Poster
Project Presentation
Koray Seyfullah DANIŞMA
Mehmet ŞEREFOĞLU
Ali Dorukhan KARAKAYA
Advisor: Dr. Engin DEMİR
Semantic Document Classification
The web is big and there are billions of text documents that cannot be searched efficiently because of the inadequacies of the current methods to classify, relate or understand the text automatically and intelligently. Most of the current efforts depend of the simple word vector representation of text documents, and since this does not involve any semantics, the classification methods are not very accurate. In this project we investigated the existing or currently being developed novel methods in semantic classification of text documents. Specifically we tested Word2Vec and Doc2Vec methods, and compared the results on two well-known datasets Reuters and BBC News. The results show that we can classify documents over 95% accuracy.
GitHub
Project Web Page
Video
Project Poster
Project Presentation
Mustafa GÜNGÖR
Merve UĞUR
Nihan GÜMÜŞ
Advisor: Prof. Dr. Erdoğan DOĞDU
Content-based Video Segmentation
A video is a sequence of images named frames. Our project aims to divide a video into segments (shots) wherein the scene or camera position does not change significantly and our project will consider both visual and semantic contents of the frames. The frames are shot in a very short time interval from each other. Therefore, the consecutive frames are very similar in content. However, with changes in scene, story, etc. the frame contents may change very quickly. For instance, a video may start with an indoor scene, and then it may switch to some outdoor scene. The first step in video indexing is video segmentation into shots where each shot is manually or automatically labeled. These labels are later used for retrieving the videos or video segments which are looked for in a query. Multimedia Computing gained traction over the last two decades which makes it fairly new. While we mention that Multimedia Computing is new, it is highly complex because it investigates human visual system, networks and devices at the same time. Our challenge is to learn current coding standards, dealing with major changes which do not necessarily correspond to the end of a video segment while minor changes in the frames can be handled easily.
GitHub
Project Web Page
Video
Project Poster
Project Presentation
Berk Can ÖZÜTEMİZ
Ece NALÇACI
Veli Engin ÖZTÜRK
Advisor: Dr. Roya CHOUPANI
Design and Implementation of a Virtual Local Area Network with using Cisco Operating System
The purpose of this document is describing the and implementation of a Virtual Local Area Network with using Cisco Operating System. This design aims to design VLANs are commonly incorporated into network design making it easier for a network to support the goals of an organization. The project will cover how to code VLANs and VLAN trunks. It will also examine security considerations and strategies relating to VLANs and trunks, and best practices for VLAN implementation and design. This document includes detailed information about requirements of the project. It reflects the identified constraints and proposed software functionalities.
GitHub
Project Web Page
Video
Project Poster
Project Presentation
Esra Nur YÜCEL
Özde ÖZAŞ
Buğra YILMAZ
Advisor: Dr. Sibel TARIYAN ÖZYER
Blood Vessel Segmentation in Eye Angiogram Images
As we progress in computer technologies, the number of solutions and software that we produce to facilitate our lives is increasing at the same rate. There has been a significant increase in the number of software solutions developed for the medical field over the past decade, especially because of the increased knowledge we have gained through studies conducted in the medical field. The goal of our work is to facilitate the ophthalmologist’s work within the overcrowded hospital because of the growing human population and therefore the increased need for medical attention. The crowded hospital creates a stress that cannot be ignored on the patients waiting in the hospital and the doctor in charge of dealing with them. For this reason, on a busy day, the time to leave for a single patient is reduced, and the doctor may see a decrease in the diagnostic accuracy due to stress. The aim of the project we have developed is to make software that will enable the ophthalmologist to make a quick and well accurate diagnosis even under the conditions mentioned above. This software is a system that can determine the locations of the retinal vessels in a fundus image and, if exists, the locations of the emboli and make an estimated diagnosis.
GitHub
Project Web Page
Video
Project Poster
Project Presentation
Aykut ER
Egeberk ÖZBERK
Advisor: Dr. Roya CHOUPANI
Biometric Identification based on Face and Iris Recognition
Today, we are living in the technology age and because of this technology it is evolving rapidly and grows day by day. Most of the people and companies are using technological products which include their privacy, so they want to keep safe their information freely. In the light of that information, some scientists have tried to produce a new product to which provide people’s private information keep in safely and avoiding the information theft. As a result of these, scientist developed Face and Iris recognition system which is the products that are the safest and easiest to use according to other security products, except those advantages, Face and Iris recognition system still has some shortcomings like illumination, external factors, so Face and Iris recognition algorithms should develop for reaching highest matching rate.
GitHub
Project Web Page
Video
Project Poster
Project Presentation
Tolga ÖZALP
Utku ÖZCAN
Edanaz PEKDEMİR
Advisor: Dr. Gül TOKDEMİR
Adroit System for Online Time Table and Announcement
The interaction between teachers and students is very significant in academic life. This interaction can become difficult. Especially, when the teachers are out of office. In these circumstances, teachers leave notes on their doors. These paper notes may get lost or get damaged and do not attract attention. With this project, these problems will be solved. Students will be able to check the teachers’ time tables, check the published announcements and send messages via using a touchscreen. On the other hand, teachers will be able to add a new announcement and edit their time tables using a mobile application or a website. Finally, admins will verify the new registered devices and add the teachers to the database. The purpose of this project is to ease the communication and interaction between teachers and students in all circumstances, prevent paper waste and design a usable and practical real-time communication system using Raspberry Pi with a touchscreen and a motion sensor, a DBMS and a web server.
GitHub
Project Web Page
Video
Project Poster
Project Presentation
Utku KILAVUZ
Özlem KILIÇ
Hatice Nazlı KUŞ
Timur ATİLA
Advisor: Dr. Abdül Kadir GÖRÜR
Curiuos Cam for Video Lecture Recording Environment
(1st place in the 12th Çankaya University R & D Project Bazaar)
Nowadays the use and importance of e-learning continue to increase, and it is well-known fact that one of the important components of e-learning is video lectures With video lectures students can view and study the instructor’s lectures as often as they wish until they understand the material. Therefore preparation of video lecture became an important issue. In its simplest form, the preparation of the video lesson requires a camera and a writing board. In order to increase the quality of video and to bring the traditional classroom environment to the video course environment, cameraman who will follow the lecturer movements is required. For some people, the presence of the cameraman can lead to stress and may degrade their teaching performance. In this respect design and implementation of a camera platform that mimics cameraman shooting techniques such as following the instructor movements will provide a more comfortable environment for video lecture recording. Therefore with this project, a microcontroller that will control camera platform using stepper-servo motors by continuously receiving inputs from motion sensors, and sound sensors is aimed to provide video lecture environment. Additionally, image processing subsystem for face recognition will also be integrated in order to let the camera to follow the instructor in case of more than one person in the shooting environment. In this document, the need analysis of the project, which has been researched in the literature survey and the differences of other projects have been taken, has been analyzed. The software process is divided into software and hardware
and the necessary materials are specified.
GitHub
Project Web Page
Video
Project Poster
Project Presentation
Bahar ŞENGEZ
Alihan Doğuş YALÇIN
Elifnur ALTUNTAŞ
Advisor: Dr. Abdül Kadir GÖRÜR
Exam Distribution System
The key problem of all schools and universities is scheduling of classes and exams. Many lecturers face with problems when they try to arrange their exam schedule. Also students may have overlaps in exam dates or exam time. With this project; we aim to develop a system which can distribute exams performed at an educational organization to the classes by arranging suitable proctors. Preventing overlaps and conflicts between any exam dates, classrooms and proctors of the exam is the main goal of this project. The system does not provide manual intervention so that overlaps and conflicts will not occur.
GitHub
Project Web Page
Video
Project Poster
Project Presentation
Başak ERGİN
Deniz ÇIKIŞ
Elif KOÇ
Advisor: Dr. Gül TOKDEMİR
Design and Implementation of Attendance Tracking and Monitoring System using BLE Beacon
In most universities, tracking the attendance in lectures is processed by sheets that include students’ id numbers, names, surnames, and signs but this method has many disadvantages. Mostly, students sign for their friends which do not attend to the lectures and if the lecturers do not count the class and compare the number ofsigns with the number ofstudents in the class, students who not attend the lectures may be calculated as they are in class. Our system is designed to track the attendance with the Bluetooth Low Energy (BLE) signals which are embedded into the BLE Beacon device and the mobile application which will be developed in Android platform and catches the BLE signals and process these signals to decide whether the studentisin the lecture or not.
GitHub
Project Web Page
Video
Project Poster
Project Presentation
Melihşah AKIN
Berkan GÜREL
Mertcan KÖKCÜR
Advisor: Dr. Murat SARAN
Smart Attendance Tracking and Monitoring System using BLE Beacon
Tracking and monitoring students’ attendance to lectures is a time-consuming activity for instructors. The proposed system automatically uses the Bluetooth Low Energy (BLE) 4.0 communication of the student smartphones as the student enters the lecture hall of the course to check the location of the beacon and automatically confirm the presence if it is valid. It is characterized by the fact that there are no actions to be taken by the students or instructors for the tracking of participation. The system also recognizes student and instructor modes with their phone numbers. This makes it more convenient to use, since there are no other actions. In addition to automatic attendance system, the project includes verified attendance system by using a mobile application that uploads the photo of the students to a server, and a detailed reporting with e-mail alert system to monitor students’ attendances. The outputs of the project are as follows: (1) Restful web services, (2) Web application, (3) iOS mobile application.
GitHub
Project Web Page
Video
Project Poster
Project Presentation
Muhammed TAŞKIN
Berkay ÇINAR
Merve ÇINAR
Advisor: Dr. Murat SARAN
Development of a Karaoke Application of Language Learning
Out-of-class learning constitutes an important context for language development. Students often have difficulty with pronunciation in English. It is thought that the karaoke offers a lot of fun and creativity while students cooperate with each other to achieve their pronunciation goals. This project is based on a speech recognition which intended to show how students can improve their pronunciation through the use of karaoke. Students will sing the song which has been determined according to their level. Interacting with songs again and again is important to language learners. By scoring their performance, it is believed that they will try to increase their score by repeating more times.
GitHub
Project Web Page
Video
Project Poster
Project Presentation
Ege Naz ALARSLAN
Mehmet Ali BEKERECİ
Tolga KARAMAN
Advisor: Dr. A. Nurdan SARAN
Story Driven, Platform and Hardware Independent Cyber Security Training Environment
Today, cyber security has become an internationally critical issue. As a result, many countries have concentrated on training cyber security professionals and in attempts to encourage people for this training, are organizing hackathons and capture the flag (CTF) competitions. These competitions run on servers and virtual machines that are specifically configured for this purpose. This project aims to create a cross-platform environment to learn and practice penetration testing skills with engaging storylines, while taking community feedback every step of the way to increase usability of the machines and increase the educational value.
GitHub
Project Web Page
Video
Project Poster
Project Presentation
Sina Barış AYDIN
Altuğ BOZKURT
Utku TORAMAN
Advisor: Dr. A. Nurdan SARAN
Autonomous Flight For Drones
An unmanned aerial vehicle (UAV), commonly known as a drone, is an aircraft without a human pilot aboard. UAVs are a component of an unmanned aircraft system (UAS); which include a UAV, a ground-based controller, and a system of communications between the two. The flight of UAVs may operate with various degrees of autonomy: either under remote control by a human operator or autonomously by onboard computers. The goal of this study is to create a autonomously flight for drone according to the coordinates entered by the user. During the flight, the drone will calculate the shortest route to the destination and will determine a new route to overcome the obstacle if there is any obstruction. Besides these, the user can specify some coordinates and the drone does not enter the forbidden zones. We will do all this in a simulation program. Ultimately, this system should create providing autonomous flight.
GitHub
Project Web Page
Video
Project Poster
Project Presentation
Selin TANKAYA
Yağmur ATICI
Advisor: Dr. Reza HASSANPOUR
Sensor Programming with using Android Platform
In this report, the features and design of a prototype robotic vehicle with using Android devices and sensors are
presented. This study was developed with the interaction of Android-based smartphone and Bluetooth technology. In addition, Arduino Uno, which is the brain of various sensors and robot, was used in this project. The aim of this project is to use for the military field to inform the military personnel in advance of possible terrorist attacks. The user can control the robotics movements from his/her own intelligent device or take the robot in automatic mode and let the car drive its own way. The robotic vehicle can detect whether the
obstacles facing the vehicle are human or not. If the obstacle is a living entity, as for that is the human, the red led on the robot lights up and it will calculate the shortest distance it can avoid and will proceed in that direction. If any object we are comparing is inanimate, it will calculate the shortest distance it can avoid and will proceed in that direction. In addition, by using sensors such as infrared sensor, obstacle recognition sensor, the interaction between the robot and the environment is ensured. At the same time, the
Bluetooth module allows Arduino to communicate with the Android-based phone. Thus, the importance of sensors is increasing day by day in the military area. Human detectable sensors are used by civilian and military personnel, especially in military areas, to ensure that necessary precautions are taken in advance to warn against possible terrorist attacks.
GitHub
Project Web Page
Video
Project Poster
Project Presentation
Esra YILMAZ
Ahmet KARA
Advisor: Dr. Sibel TARIYAN ÖZYER
Design and Implementation of a Sensor Network
While technology is evolving very quickly, many innovations and facilities has brought. In this respect, technology has become an indispensable part of our life. One of the innovations is sensor technology. Today, using this system technology; It is used frequently in factories, parking lots, libraries, cars, museums. Thanks to the innovations brought by these sensors, technological products enter into human life more actively and provide a lot of convenience. The great benefits are that they gain a positive contribution to the environment by gaining time and work power. we use the LDR sensor, which is the sensor we use, to save electricity.
GitHub
Project Web Page
Video
Project Poster
Project Presentation
Taylan ERSEN
Haluk Mesut MERCİMEK
Buse BİÇER
Advisor: Dr. Sibel TARIYAN ÖZYER
